You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If the hard coded batch size can instead be edited in the advanced config that would be great. I figured out, after some testing that the batch size is hard coded in app.py file. I had initial tried changing the batch size via the advanced commands, but noticed it didn't increase vram usage or change the speed of the training.
So to actually change the batch size I:
Tried the advanced command "--train_batch_size", which doesn't work since its overridden by the hard coded batch size
Edited the hard coded batch size to what I needed, and noticed it changed the training parameters and actually worked (used more vram & visible difference in training speed).
The text was updated successfully, but these errors were encountered:
If the hard coded batch size can instead be edited in the advanced config that would be great. I figured out, after some testing that the batch size is hard coded in app.py file. I had initial tried changing the batch size via the advanced commands, but noticed it didn't increase vram usage or change the speed of the training.
So to actually change the batch size I:
The text was updated successfully, but these errors were encountered: